Exposure Data Test

GEOG 394 Exposures Across Space and Time

Author

Alayna Johnson

Published

February 11, 2025

Loading Dependencies

library(tidyverse)
library(ggplot2)
library(plotly)
library(sf)
library(leaflet)
library(viridis)

Reading in Data

grandOlri <- read_csv("1505_Grand_Ave-2025-11-2_6-49-09 PM.csv")

portlandSaratoga <- read_csv("1760_Portland_Ave-2025-11-2_9-42-04 PM.csv")

Plotting Accelerometer Data

Accelerometer Measuring

Left: iOS device acceleration directions, right: adjusted acceleration directions

The measure that we are most interested in when looking at road quality is the upward and downward acceleration of the device mounted on the vehicle. iOS devices measure accelerations in the directions seen in the image on the left. Intuitively, it makes more sense for upwards Y acceleration to be positive and downwards Y to be negative as well as right X as positive and left X as negative. These adjustments were made to the data to make plots more intuitive. The image on the right shows what the directions of acceleration measurement look like now.

The iOS device used to collect these accelerometer data is mounted on the inside of the vehicle in a vent phone mount. This means that the device sits straight up and down like the images above.

Drive from Portland to Saratoga

This drive starts on Portland and Wheeler going through Summit Ave, Fairview Ave, Grand Ave, and ending on Saratoga.

#| cache: true

# pivot_longer where accelerometer axis measurements are all in one col instead of separate
portlandSaratoga_long <- portlandSaratoga %>%
  mutate(accelerometer_y_new = -1* accelerometer_y,
         accelerometer_x_new = -1 * accelerometer_x) %>%
  pivot_longer(cols = c(accelerometer_x_new, accelerometer_y_new, accelerometer_z),
               names_to = "axis", values_to = "acceleration") %>%
  mutate(
    axis = recode(axis, 
                  "accelerometer_x_new" = "X", 
                  "accelerometer_y_new" = "Y", 
                  "accelerometer_z" = "Z"),
     alpha_value = ifelse(axis == "Y", 1, 0.3)
  )

# mutate column to just have time for X axis
portlandSaratoga_long <- portlandSaratoga_long %>%
  group_by(accelerometer_local_time) %>%
  mutate(
    time_label = ifelse(axis == "X", paste("Time:", accelerometer_local_time), "")
  ) %>%
  ungroup()

# plot
p <- ggplot(portlandSaratoga_long, aes(x = accelerometer_local_time, y = acceleration, color = axis, group = axis)) +
  geom_line(aes(text = paste(time_label, 
                             "<br>Axis:", axis, 
                             "<br>Acceleration:", acceleration),
                alpha = alpha_value)) +
  labs(title = "Acceleration - Portland Ave to Saratoga St", x = "Time", y = "Acceleration", color = "Axis") +
  theme_minimal() +
  scale_alpha_identity()


p_plotly <- ggplotly(p, tooltip = "text") %>%
  layout(
    hovermode = "x unified",  # adding trace line/crosshair
    shapes = list(
      list(
        type = "line",
        xref = "x",
        yref = "paper",
        x0 = 0, x1 = 0,  
        y0 = 0, y1 = 1,  
        line = list(color = "darkgray", width = 1, dash = "dash")  # Dashed vertical line
      )
    )
  )

p_plotly

Drive from Saratoga to Olri

This drive starts on Saratoga and Summit, drives through Snelling, Grand, and Macalester St ending in the parking lot by the tennis courts.

# pivot_longer where accelerometer axis measurements are all in one col instead of separate
grandOlri_long <- grandOlri %>%
 mutate(accelerometer_y_new = -1* accelerometer_y,
         accelerometer_x_new = -1 * accelerometer_x) %>%
  pivot_longer(cols = c(accelerometer_x_new, accelerometer_y_new, accelerometer_z),
               names_to = "axis", values_to = "acceleration") %>%
  mutate(
    axis = recode(axis, 
                  "accelerometer_x_new" = "X", 
                  "accelerometer_y_new" = "Y", 
                  "accelerometer_z" = "Z"),
     alpha_value = ifelse(axis == "Y", 1, 0.3)
  )

# mutate column to just have time for X axis
grandOlri_long <- grandOlri_long %>%
  group_by(accelerometer_local_time) %>%
  mutate(
    time_label = ifelse(axis == "X", paste("Time:", accelerometer_local_time), "")
  ) %>%
  ungroup()

# plot
p2 <- ggplot(grandOlri_long, aes(x = accelerometer_local_time, y = acceleration, color = axis, group = axis)) +
  geom_line(aes(text = paste(time_label, 
                             "<br>Axis:", axis, 
                             "<br>Acceleration:", acceleration),
                alpha = alpha_value)) +
  labs(title = "Acceleration - Saratoga St to Macalester St", x = "Time", y = "Acceleration", color = "Axis") +
  theme_minimal() +
  scale_alpha_identity()


p_plotly2 <- ggplotly(p2, tooltip = "text") %>%
  layout(
    hovermode = "x unified",  # adding trace line/crosshair
    shapes = list(
      list(
        type = "line",
        xref = "x",
        yref = "paper",
        x0 = 0, x1 = 0,  
        y0 = 0, y1 = 1,  
        line = list(color = "darkgray", width = 1, dash = "dash")  # Dashed vertical line
      )
    )
  )

p_plotly2

Mapping Accelerometer Data

combined <- rbind(grandOlri, portlandSaratoga) %>%
  mutate(accelerometer_y_new = -1 * accelerometer_y,
         accelerometer_x_new = -1 * accelerometer_x) %>%
  filter(!is.na(location_longitude)& !is.na(location_latitude)) %>%
  select(-location_altitude, -location_speedAccuracy, -location_bearingAccuracy, - location_altitudeAboveMeanSeaLevel, -location_bearing, -location_horizontalAccuracy, location_verticalAccuracy, -location_speed)

# changing csv to sf object for mapping, EPSG:4326
combined_sf <- st_as_sf(combined, coords = c("location_longitude", "location_latitude"), crs = 4326)

# creating continuous color palette with turbo color scale from viridis
pal <- colorNumeric(palette = rev(viridis::turbo(100)), domain = combined_sf$accelerometer_y_new)

# Create Leaflet map
leaflet(combined_sf) %>%
  addProviderTiles("CartoDB.Positron") %>%
  addCircles(
    radius = 1,
    color = ~pal(accelerometer_y_new),  # continuous color scale for y acceleration
    weight = 5,
    opacity = 0.9,
    popup = ~paste0("Acceleration Y: ", accelerometer_y_new)
  ) %>%
  addLegend(
    pal = pal, 
    values = combined_sf$accelerometer_y_new, 
    title = "Accelerometer Y",
    opacity = 1
  )

comparing times –> subtracting one from the other to see change over time

rasterize? how?????????????????
projecting the points onto a line on each side of the street?
interpolation for points, each second of measurement
Toggle for continuous vs discrete colors of y acceleration (discrete being only the worst values, red for down and blue for up)
Starting by driving both directions on main roads: Summit, Grand, Snelling, Fairview, Saint Clair
going into the neighborhoods, do I really want to go all the way to Saratoga?? Prob not since I have to do this drive multiple times…

--- title: "Exposure Data Test" subtitle: "GEOG 394 Exposures Across Space and Time" author: "Alayna Johnson" date: "February 11, 2025" format: html: toc: true toc-depth: 3 embed-resources: true code-tools: true --- ```{r setup, include=FALSE} knitr::opts_chunk$set(echo = TRUE, error = TRUE, warning = FALSE) ``` # Loading Dependencies ```{r loading-packages} #| message: false library(tidyverse) library(ggplot2) library(plotly) library(sf) library(leaflet) library(viridis) ``` ### Reading in Data ```{r reading-data} #| message: false grandOlri <- read_csv("1505_Grand_Ave-2025-11-2_6-49-09 PM.csv") portlandSaratoga <- read_csv("1760_Portland_Ave-2025-11-2_9-42-04 PM.csv") ``` # Plotting Accelerometer Data ## Accelerometer Measuring ![Left: iOS device acceleration directions, right: adjusted acceleration directions](accelerometer-axes.png){width=70%} The measure that we are most interested in when looking at road quality is the upward and downward acceleration of the device mounted on the vehicle. iOS devices measure accelerations in the directions seen in the image on the left. Intuitively, it makes more sense for upwards Y acceleration to be positive and downwards Y to be negative as well as right X as positive and left X as negative. These adjustments were made to the data to make plots more intuitive. The image on the right shows what the directions of acceleration measurement look like now. The iOS device used to collect these accelerometer data is mounted on the inside of the vehicle in a vent phone mount. This means that the device sits straight up and down like the images above. ## Drive from Portland to Saratoga This drive starts on Portland and Wheeler going through Summit Ave, Fairview Ave, Grand Ave, and ending on Saratoga. ```{r plotting-accelerometer-home} #| cache: true # pivot_longer where accelerometer axis measurements are all in one col instead of separate portlandSaratoga_long <- portlandSaratoga %>% mutate(accelerometer_y_new = -1* accelerometer_y, accelerometer_x_new = -1 * accelerometer_x) %>% pivot_longer(cols = c(accelerometer_x_new, accelerometer_y_new, accelerometer_z), names_to = "axis", values_to = "acceleration") %>% mutate( axis = recode(axis, "accelerometer_x_new" = "X", "accelerometer_y_new" = "Y", "accelerometer_z" = "Z"), alpha_value = ifelse(axis == "Y", 1, 0.3) ) # mutate column to just have time for X axis portlandSaratoga_long <- portlandSaratoga_long %>% group_by(accelerometer_local_time) %>% mutate( time_label = ifelse(axis == "X", paste("Time:", accelerometer_local_time), "") ) %>% ungroup() # plot p <- ggplot(portlandSaratoga_long, aes(x = accelerometer_local_time, y = acceleration, color = axis, group = axis)) + geom_line(aes(text = paste(time_label, "<br>Axis:", axis, "<br>Acceleration:", acceleration), alpha = alpha_value)) + labs(title = "Acceleration - Portland Ave to Saratoga St", x = "Time", y = "Acceleration", color = "Axis") + theme_minimal() + scale_alpha_identity() p_plotly <- ggplotly(p, tooltip = "text") %>% layout( hovermode = "x unified", # adding trace line/crosshair shapes = list( list( type = "line", xref = "x", yref = "paper", x0 = 0, x1 = 0, y0 = 0, y1 = 1, line = list(color = "darkgray", width = 1, dash = "dash") # Dashed vertical line ) ) ) p_plotly ``` ## Drive from Saratoga to Olri This drive starts on Saratoga and Summit, drives through Snelling, Grand, and Macalester St ending in the parking lot by the tennis courts. ```{r plotting-accelerometer-olri} #| cache: true # pivot_longer where accelerometer axis measurements are all in one col instead of separate grandOlri_long <- grandOlri %>% mutate(accelerometer_y_new = -1* accelerometer_y, accelerometer_x_new = -1 * accelerometer_x) %>% pivot_longer(cols = c(accelerometer_x_new, accelerometer_y_new, accelerometer_z), names_to = "axis", values_to = "acceleration") %>% mutate( axis = recode(axis, "accelerometer_x_new" = "X", "accelerometer_y_new" = "Y", "accelerometer_z" = "Z"), alpha_value = ifelse(axis == "Y", 1, 0.3) ) # mutate column to just have time for X axis grandOlri_long <- grandOlri_long %>% group_by(accelerometer_local_time) %>% mutate( time_label = ifelse(axis == "X", paste("Time:", accelerometer_local_time), "") ) %>% ungroup() # plot p2 <- ggplot(grandOlri_long, aes(x = accelerometer_local_time, y = acceleration, color = axis, group = axis)) + geom_line(aes(text = paste(time_label, "<br>Axis:", axis, "<br>Acceleration:", acceleration), alpha = alpha_value)) + labs(title = "Acceleration - Saratoga St to Macalester St", x = "Time", y = "Acceleration", color = "Axis") + theme_minimal() + scale_alpha_identity() p_plotly2 <- ggplotly(p2, tooltip = "text") %>% layout( hovermode = "x unified", # adding trace line/crosshair shapes = list( list( type = "line", xref = "x", yref = "paper", x0 = 0, x1 = 0, y0 = 0, y1 = 1, line = list(color = "darkgray", width = 1, dash = "dash") # Dashed vertical line ) ) ) p_plotly2 ``` # Mapping Accelerometer Data ```{r combined-data} combined <- rbind(grandOlri, portlandSaratoga) %>% mutate(accelerometer_y_new = -1 * accelerometer_y, accelerometer_x_new = -1 * accelerometer_x) %>% filter(!is.na(location_longitude)& !is.na(location_latitude)) %>% select(-location_altitude, -location_speedAccuracy, -location_bearingAccuracy, - location_altitudeAboveMeanSeaLevel, -location_bearing, -location_horizontalAccuracy, location_verticalAccuracy, -location_speed) ``` ```{r initial-map} #| cache: true # changing csv to sf object for mapping, EPSG:4326 combined_sf <- st_as_sf(combined, coords = c("location_longitude", "location_latitude"), crs = 4326) # creating continuous color palette with turbo color scale from viridis pal <- colorNumeric(palette = rev(viridis::turbo(100)), domain = combined_sf$accelerometer_y_new) # Create Leaflet map leaflet(combined_sf) %>% addProviderTiles("CartoDB.Positron") %>% addCircles( radius = 1, color = ~pal(accelerometer_y_new), # continuous color scale for y acceleration weight = 5, opacity = 0.9, popup = ~paste0("Acceleration Y: ", accelerometer_y_new) ) %>% addLegend( pal = pal, values = combined_sf$accelerometer_y_new, title = "Accelerometer Y", opacity = 1 ) ``` \ \ comparing times --> subtracting one from the other to see change over time - rasterize? how????????????????? - projecting the points onto a line on each side of the street? - interpolation for points, each second of measurement - Toggle for continuous vs discrete colors of y acceleration (discrete being only the worst values, red for down and blue for up) - Starting by driving both directions on main roads: Summit, Grand, Snelling, Fairview, Saint Clair - going into the neighborhoods, do I really want to go all the way to Saratoga?? Prob not since I have to do this drive multiple times...